Technology-aware multi-domain multi-layer routing

IEEE INFOCOM 2016 - The 35th Annual IEEE International Conference on Computer Communications

Fergani

Cohen

et al. 2016

Abstract-Carrier-grade networks comprise several layers where different protocols coexist. Nowadays, most of these networks have different control planes to manage routing on different layers, leading to a suboptimal use of the network resources and additional operational costs. However, some routers are able to encapsulate, decapsulate and convert protocols and act as a liaison between these layers. A unified control plane would be useful to optimize the use of the network resources and automate the routing configurations. Software-Defined Networking (SDN) based architectures, such as OpenFlow, offer a chance to design such a control plane. One of the most important problems to deal with in this design is the path computation process. Classical path computation algorithms cannot resolve the problem as they do not take into account encapsulations and conversions of protocols. In this paper, we propose algorithms to solve this problem and study several cases: Path computation without bandwidth constraint, under bandwidth constraint and under other Quality of Service constraints. We study the complexity and the scalability of our algorithms and evaluate their performances on real topologies. The results show that they outperform the previous ones proposed in the literature.

Section: ) Network Used For the Simulations And Methodologymentioning

confidence: 99%

Path computation in multi-layer networks: Complexity and algorithms

IEEE INFOCOM 2016 - The 35th Annual IEEE International Conference on Computer Communications

Fergani

Cohen

et al. 2016

“…In [3], they widely generalize the previous algorithms, and prove that the feasible path problem under bandwidth constraint remains NPhard even with only two protocols and on symmetric directed graphs. The authors of [8] propose a matrix based model and an algorithm to compute feasible paths. However, the algorithm takes exponential time, and can only compute loopless feasible paths.…”

Section: A Problem Definition and Illustrationmentioning

confidence: 99%

A stack-vector routing protocol for automatic tunneling

IEEE INFOCOM 2019 - IEEE Conference on Computer Communications

Lassourreuille

Kunne

et al. 2019

In a network, a tunnel is a part of a path where a protocol is encapsulated in another one. A tunnel starts with an encapsulation and ends with the corresponding decapsulation. Several tunnels can be nested at some stage, forming a protocol stack. Tunneling is very important nowadays and it is involved in several tasks: IPv4/IPv6 transition, VPNs, security (IPsec, onion routing), etc. However, tunnel establishment is mainly performed manually or by script, which present obvious scalability issues. Some works attempt to automate a part of the process (e.g., TSP, ISATAP, etc.). However, the determination of the tunnel(s) endpoints is not fully automated, especially in the case of an arbitrary number of nested tunnels. The lack of routing protocols performing automatic tunneling is due to the unavailability of path computation algorithms taking into account encapsulations and decapsulations. There is a polynomial centralized algorithm to perform the task. However, to the best of our knowledge, no fully distributed path computation algorithm is known. Here, we propose the first fully distributed algorithm for path computation with automatic tunneling, i.e., taking into account encapsulation, decapsulation and conversion of protocols. Our algorithm is a generalization of the distributed Bellman-Ford algorithm, where the distance vector is replaced by a protocol stack vector. This allows to know how to route a packet with some protocol stack. We prove that the messages size of our algorithm is polynomial, even if the shortest path can be of exponential length. We also prove that the algorithm converges after a polynomial number of steps in a synchronized setting. We adapt our algorithm into a proto-protocol for routing with automatic tunneling and we show its efficiency through simulations.

“…They used automata and language theory tools to compute a shortest feasible path, but only considering the number of hops or adaptation functions. More recently, Iqbal et al [21] underlined the need of path computation algorithms in NRENs. They proposed a new matrix-based model for multi-layer networks and k-shortest paths algorithms to tackle the problem.…”

Section: B Related Workmentioning

confidence: 99%

“…However, this probability quickly decreases. The phase transition phenomenon can be seen in [21]. But the results consider only loopless paths and deals with technologies rather than adaptation functions.…”

Section: A Path Computation Without Bandwidth Constraintmentioning

confidence: 99%

Algorithmic and Complexity Aspects of Path Computation in Multi-Layer Networks

IEEE/ACM Trans. Networking

Fergani

Cohen

2018

Carrier-grade networks comprise several layers where different protocols coexist. Nowadays, most of these networks have different control planes to manage routing on different layers, leading to a suboptimal use of the network resources and to additional operational costs. However, some routers are able to encapsulate, decapsulate, and convert protocols, and act as a liaison between these layers. A unified control plane would be useful to optimize the use of the network resources and to automate the routing configurations. Software-Defined Networking based architectures, offer an opportunity to design such a control plane. One of the most important problems to deal with in this design is the path computation process. Classical path computation algorithms cannot resolve the problem as they do not take into account encapsulations and conversions of protocols. In this paper, we propose algorithms to solve this problem, and we study several cases. If there is no bandwidth constraint, we propose a polynomial algorithm that compute the optimal path. We also give lower and upper bounds on the optimal path length. On the other hand, we show that the problem is NPhard if there is a bandwidth constraint (or other Quality of Service parameters), even if there is only two protocols and in a symmetric graph. We study the complexity and the scalability of our algorithms and evaluate their performances on real and random topologies. The results show that they are faster than the previous ones proposed in the literature. These algorithms can also have important applications in automatic tunneling.